Why Public Works Are Poor Stimulus and Why We Still Need Them
For some odd reason, it seems to be news that the public works projects financed by the 2009 stimulus bill have been very slow to come online because there were few shovel-ready projects in the pipeline that needed only funding to get started. Everyone appears to be treating this statement by President Obama in today’s New York Times Magazine as a revelation: “There’s no such thing as shovel-ready projects.”
Indeed, some are reacting as if it is a scandal that Obama apparently had this insight a year ago and disclosed it off-the-record to Times columnist David Brooks.
If this is a scandal, then it may be the most idiotic one I have seen in a long time. Of course there were few if any shovel-ready projects. Economists have known this for decades and it is an important reason why public works are a poor means of stimulating short-run growth. By the time the projects come online most recessions are long over. Here is what I wrote in The Public Interest almost 20 years ago, summarizing a 1980 Office of Management and Budget study of countercyclical public works:
● Public works programs cannot be triggered and targeted in a sufficiently timely manner to compensate for cyclical fluctuations in unemployment and economic activity.
● Even if it were possible to properly time a countercyclical program, the time it takes to construct public works would lead to a significant overlap of job generation and economic stimulus with periods of economic recovery.
● Public works programs have had a minimal impact on the unemployed. This is partly because the programs are not labor-intensive, and partly because many of the jobs created require skills the unemployed do not have.
● The duration of employment for individual workers is too short to provide meaningful economic relief, to maintain skills and work habits, or to provide on-the-job training.
● Public works are extremely costly. The cost of generating a construction job for one year ranges from $70,000 to $198,000. [Note: these are 1980 dollars; adjusted for inflation, it would be $185,000 to $525,000 per job.]
These conclusions weren’t even new in 1980. The long lead time for public works startups had been documented in the academic literature long before that. See, for example, Sherman Maisel, “Timing and Flexibility of a Public Works Program,” Review of Economics and Statistics (May 1949), pp. 147-152. Indeed, John Maynard Keynes himself expressed skepticism about using public works for countercyclical purposes. As he wrote in 1942:
Organised public works, at home and abroad, may be the right cure for a chronic tendency to a deficiency of effective demand. But they are not capable of sufficiently rapid organization (and above all cannot be reversed or undone at a later date), to be the most serviceable instrument for the prevention of the trade cycle. [Collected Writings, vol. 27, p. 122.]
Because I knew the history of public works and countercyclical policy, I warned anyone who would listen that it was going to take a long time for public works programs to stimulate growth and reduce unemployment. Here’s what I wrote in a January 9, 2009 Forbes column:
Of course, Obama isn't only relying on tax policy to stimulate growth; he has plans to increase spending on public works as well. But this may not do much to create jobs this year. Despite claims by the Conference of Mayors and the transportation lobby that there is as much as $96 billion in construction "ready to go," the fact is that it takes a long time before meaningful numbers of workers can be hired for such projects.
As a recent Congressional Budget Office study explains, "Practically speaking ... public works involve long start-up lags. ... Even those that are 'on the shelf' generally cannot be undertaken quickly enough to provide timely stimulus to the economy."
Here’s what I wrote in a January 23, 2009 column:
The problem is that fiscal stimulus needs to be injected right now to counter the liquidity trap. If that were the case, I think we might well get a very high multiplier effect this year. But if much of the stimulus doesn't come online until next year, when we are likely to be past the worst of the slowdown, then crowding out will greatly diminish the effectiveness of the stimulus, just as the critics argue. According to the Congressional Budget Office, only a fraction of proposed infrastructure spending can be spent before October of next year; the bulk would come long after.
And here is what I said in a June 26, 2009 column:
The problem is that Obama was always much too optimistic about how quickly stimulus spending could have an effect. As I warned in a January column, it takes far more time for it to impact the economy than most people think. Moreover, not all government spending is necessarily stimulative, and the parts of the stimulus package that provide real stimulus are among the slowest to come online.
Congressional Budget Office Director Douglas Elmendorf recently presented a report to the International Monetary Fund in which he walked through some of the problems with implementing the stimulus program.
First of all, 60% of the stimulus package was never going to have much of a stimulative effect. These were programs like extending unemployment benefits and tax credits with no incentive effects that may have been justified on the merits, but don't really do anything to increase growth or reduce unemployment.
For a program to be stimulative, it must bring forth economic activity that otherwise would not have taken place. The classic example is public works. When a new road or bridge is built, construction companies have to purchase concrete, steel and other materials that create business for other companies. They also employ workers that otherwise would not be working, paying them wages that they will spend, producing jobs and incomes for other workers.
If this works the way it is supposed to, stimulus spending has a multiplier effect throughout the economy. A Council of Economic Advisers study estimated that government purchases of goods and services raise the gross domestic product by $1.57 for every $1 spent. By contrast, tax credits and income transfers are much less stimulative, raising GDP by considerably less than $1 for every $1 rise in the deficit.
Since 60% of the stimulus package had a multiplier effect of less than one, only 40% of the package went to programs like public works that have a high multiplier. Moreover, the programs with a low multiplier were the fastest ones to implement; those with a high multiplier take much more time to come online. According to Elmendorf, by the end of fiscal year 2009, which ends on Sept. 30, about a third of the least stimulative spending will have been spent vs. only 11% of the highly stimulative spending.
Even at the end of fiscal year 2010, we will have spent only 47% of the highly stimulative spending. By the end of fiscal year 2011, more than a quarter of the stimulative spending will still remain unspent.
The CBO bases this estimate on many years of experience with various types of government programs. Increases in defense spending are the quickest to stimulate because 65% of the money is usually spent in the first year, rising to 88% the second year and 96% in the third. By contrast, only 27% of highway spending is spent the first year, rising to 68% the second year and 84% the third. Spending on water projects is even slower to come online, with only 4% spent the first year, rising to 24% the second year and 54% the third.
The reason is simple. Much of what the Defense Department buys involves things that also have civilian uses. In effect, it is buying off the shelf so spending can be done quickly. But to build a new highway or dam is much more complicated. Plans have to be drawn up, land acquired, environmental impact statements prepared, public comments solicited, political and other objections dealt with, contracts written and put out for bid, etc. This takes years and years.
Of course, there were a few shovel-ready projects for which all the preliminary work had been done that only needed money to start work. But there were far fewer of these projects than generally believed. Moreover, as Popular Mechanics reported, many of those that were ready to go were those that had been shelved because they were outdated or had other flaws.
Even the simplest public works projects such as road repaving take months to get moving from the time a federal check arrives. And in the short run such small projects have very little stimulative potential because state and local governments will simply use the workers and materials they already have on hand to do the work.
In conclusion, it is surprising that the lack of shovel-ready projects or the slow pace of public works spending is a revelation to anyone. Obama’s statement is simply an admission of the obvious. I have no doubt that Larry Summers, Christie Romer and Peter Orszag all told him that public works were not likely to provide much short-term stimulus. Their main value is that public goods stimulate long-term growth and the best time to produce them is when the private sector has a lot of idle resources and they can be financed at low interest rates. These are still very good reasons to undertake public works now before the economy recovers, interest rates rise and the private sector competes for the resources necessary for large public works projects.